Introduction: The presence of axillary lymph node involvement is an important prognostic factor and has a major impact on treatment decisions in early breast cancer patients. This study aimed to determine the role of cortical thickness in axillary ultrasound (AUS) as an indicator of lymph node metastasis. Methods: 766 patients with primary breast cancer who received AUS during clinical work-up were selected for this retrospective study. Lymph nodes were defined as suspicious if they showed a cortical thickness of > 3,0 mm at 11-15 MHz harmonic imaging ultrasound. Lymph node involvement was assessed by core needle biopsy (n=150), sentinel node dissection or axillary dissection. Extensive axillary spread (EAS) was diagnosed if more than two lymph nodes showed metastatic disease on histology. Results: AUS for detecting all lymph node metastases had a sensitivity of 62,27 %, a specificity of 93,15 % and a negative predictive value of 81,74 %. However, the resulting negative predictive value for transcapsular growth was 93,97 %, and for extensive axillary spread 97,52 %. Conclusion: Extensive axillary spread - in contrast to non-palpable involvement of 1 or 2 lymph nodes - contributes relevantly to the individualization of breast cancer treatment. In combination with SNB, axillary ultrasound using cortical thickness as the main distinctive parameter seems to be an easily available, robust tool of diagnosing extensive axillary metastases. If axillary ultrasound proves negative, it helps to reduce the number of classic axillary dissections.